Passive Device-free Indoor Movement Detection using WiFi Time-of-Flight Information

We present ToFMD, a system that can passively detect the movement of an indoor device-free user by relying on the absorptions and reflections of WiFi signals off his body. Our system is of practical use as it is able to reuse existing deployed commodity 802.11 infrastructure networks. It is based on the observation that the movement of an user is altering the multipath characteristics of the wireless channel which leads to changes in the measured time-of-flight (ToF) between adjacent WiFi access points. The system can detect moving people at different places in the room, it is not necessary that the user blocks the line-of-sight (LOS) path. In contrast to other approaches, our system works with inexpensive single antenna WiFi systems without access to low-level parameters like channel state information (CSI). Instead the ToF between any pair of access points is estimated using the fine time measurement (FTM) protocol, which is standardized within IEEE 802.11. The movement is detected with the help of a one class support vector machine. As proof of concept, we implemented our system with ultra-low cost ESP32 microcontrollers. The evaluation shows that our solution outperforms traditional approaches using RSSI except for the scenario with user motion in adjacent rooms. If our system is installed in the same room as the user, it is able to detect its movement with an accuracy of at least 95 %.